Complex operational systems are subject to unexpected changes in condition over time. As a side effect of their operation, these systems produce multiple signals that can be captured. These signals may include currents, temperatures, positions, forces, material properties, etc. Patterns present in these signals may be used to understand the current or future condition of a complex operational system as it is operating. For example, a stamping machine that produces steel parts will periodically enter a condition in which maintenance is required. During operation, measurements of currents, forces, and temperatures are captured, and can be used to identify the onset of a maintenance-needed condition.